Access to Greenspace in Worcester

(Note: This GIS analysis was completed in the second half of 2020 as part of a social applications of GIS course)

Figure 1

Figure 2

Q1.

How many individual recreation areas are in this dataset? List three largest places.

There are 1,513 recreation areas in the dataset.

The three largest places are:

· Leominster State Forest

· Upton State Forest

· George H. Nichols Flood Control Site

Q2.

How many records are there in the attribute table of this layer?

There are 52 records in the attribute table of this layer.

Q3.

How many records are in this attribute table? How does this compare to your answer in Q2?

There are 2,968 records in this attribute table. This number is higher than the number of records in Q2 — the residential areas have been separated out from single part to multi part polygons.

Q4.

What are the shortest and the longest distances (excluding values of 0)? In what parts of Worcester are they?

Total length field gives average distance in meters from all residential areas within each census tract to the nearest recreational space.

Q.5

What is the min, average and max value for the mean distance for census tracts to the nearest rec space? Are these values different from the answer to Q4? Explain why or why not.

· Maximum: 1,032.02 meters

· Average: 634.58 meters

· Minimum: 127.05 meters

These values are different from the answer in Q4 because they are the minimum, maximum, and average of the mean distance for census tracts to the nearest recreational space. The values in Q4 was the minimum and maximum distance between one census tract

Q6.

What classification did you use for your bivariate maps? Are your two maps similar or different? What story can you tell about the access to rec space using these maps? Your story may consider these questions from the beginning of the tutorial — how close are parks to residential areas in Worcester? How far do people with no access to a personal vehicle have to walk to the nearest recreational space? What areas are disadvantaged in terms of access to a recreational space?

The classification I used for my bivariate maps is the equal intervals classification. The two maps are similar — “Number of households with no vehicle available” and “percent of households with no vehicle available” mapped using an equal intervals classification ends up having similar equal intervals.

Overall, census tracts at the centre of Worcester have the highest average distance from green spaces and the highest number and percentage of households with no car. The census tracts that are closest to greenspaces and have the highest number and percentage of households with a vehicle are in the south east of Worcester.

On average across Worcester, people with no access to a personal vehicle walk 638.54 meters to the nearest recreational space. Those who walk the least to access a recreational space walk on average 127.05 meters. Those who walk the most walk on average 1,032.02 meters.

Q7.

In this exercise we used a combination of advanced vector analysis techniques — asymmetric mapping (using land cover to extract residential areas within census tracts) and network analysis. We made several decisions along the way — about what to include as recreational space, and how to represent census tracts and rec space in the network analysis. Review the steps we took and write a 100-word reflection on the methods used here. Now, how can this analysis be refined?

For the network analysis conducted — in this case — the incidents refer to the residential areas within census tracts and facilities refer to recreational spaces. The incidents within each census tract were clipped to the census tract. The recreational spaces within each census tract were also clipped to the census tract. Centroids were made for residential areas, and centroids were also made for recreational spaces. Based on these centroids. The average distance of residential areas to recreational spaces were calculated for each census tract.

However, centroids for large recreational areas may overestimate the distance of residential areas to recreational centres. Recreational areas do not have equal length and breath. Centroids of recreational areas that are long, for example, would create the impression that residential areas (centroids) are far away. In cases where necessary, this analysis can use the outer perimeter of recreational spaces (facilities) and residential areas (incidents) and not their centroids.

Also, the use of average distance may create underestimates or overestimates of distance and may make the output of the network analysis not comparable across census tracts. Some census tracks might have residential areas that are outliers that shoot average distance up. Residential areas may be clustered and densely populated in another area and depending on the number of recreational spaces in that census tract, the average distance could be a smaller number. In cases where appropriate, a measure of dispersion of incidents (residential areas) may need to be accounted for.